New algorithm for integration between wireless microwave sensor network and radar for improved rainfall measurement and mapping

Abstract

Abstract. One of the main challenges for meteorological and hydrological modelling is accurate rainfall measurement and mapping across time and space. To date, the most effective methods for large-scale rainfall estimates are radar, satellites, and, more recently, received signal level (RSL) measurements derived from commercial microwave networks (CMNs). While these methods provide improved spatial resolution over traditional rain gauges, they have their limitations as well. For example, wireless CMNs, which are comprised of microwave links (ML), are dependant upon existing infrastructure and the ML' arbitrary distribution in space. Radar, on the other hand, is known in its limitation for accurately estimating rainfall in urban regions, clutter areas and distant locations. In this paper the pros and cons of the radar and ML methods are considered in order to develop a new algorithm for improving rainfall measurement and mapping, which is based on data fusion of the different sources. The integration is based on an optimal weighted average of the two data sets, taking into account location, number of links, rainfall intensity and time step. Our results indicate that, by using the proposed new method, we not only generate more accurate 2-D rainfall reconstructions, compared with actual rain intensities in space, but also the reconstructed maps are extended to the maximum coverage area. By inspecting three significant rain events, we show that our method outperforms CMNs or the radar alone in rain rate estimation, almost uniformly, both for instantaneous spatial measurements, as well as in calculating total accumulated rainfall. These new improved 2-D rainfall maps, as well as the accurate rainfall measurements over large areas at sub-hourly timescales, will allow for improved understanding, initialization, and calibration of hydrological and meteorological models mainly necessary for water resource management and planning.

Multimodal Data Fusion: An Overview of Methods, Challenges, and Prospects

“…Rain gauges, as an example, are simply cups that collect the precipitation. Albeit the most direct and reliable technique, their small sampling area implies very localized representativeness and thus poor spatial resolution (e.g., [46] and [47]). Rain gauges may be read automatically at intervals as short as seconds.…”

“…Microwave links are deployed by cellular providers for backhaul communication between base stations. The signals transmitted by the base stations are influenced by various atmospheric phenomena (e.g., [49]), primarily attenuation due to rainfall [46], [47]. These changes in signal strength are recorded at predefined time intervals and kept in the cellular provider's logs.…”

“…Altogether, these technologies are largely complementary in their ability to detect and distinguish between different meteorological phenomena, spatial coverage, temporal resolution, measurement error, and other properties. Therefore, meteorological data are often combined for better accuracy, coverage and resolution; see, e.g., [19], [47], [48], and references therein.…”

“…An audio signal is usually sampled at several kilohertz whereas the video signal is typically sampled at 15-60 Hz [144] (Example II-A.1). In meteorological monitoring (Example II-C.2), each modality has very distinct spatial and temporal resolutions; this is probably the reason why solutions based on data integration [47] (see Section V-A) are preferred. Different sampling schemes in coupled matrix and tensor decompositions are discussed, e.g., in [145] and [146].…”

“…Similar to [153], these weights are interpreted as noise variances. Finally, in [47], Liberman et al process several meteorological monitoring modalities separately and then make a soft decision, using an optimal weighted average based on location, number of links, rainfall intensity and other parameters.…”